Custom Metrics

We make it possible for enterprise-level organizatons to send custom metrics to VividCortex in an easy way. Our agent is a StatsD-like collector: you can send metrics to it in the form of a short message. To get started, contact Customer Support and we will provide additional details and assistance.

The message format is very similar to a standard StatsD client:

({hostID}) {metric name} {metric value} {metric type initial}

The hostID is optional and usually it should not be included as the agent will do so automatically, the exception being when you’re collecting a metric on behalf of a different server. You can find the hostID for a given server in the Inventory page, hovering the mouse pointer over the hostname of the server; it is an integer.

Metrics you send to VividCortex should have a dot-separated name, which defines it within a namespace hierarchy or tree. Metric names should only include lowercase letters, numbers, underscores, and dots. They cannot be composed of numbers alone. In order to prevent possible collisions, please prefix your custom metrics with either ‘custom’ or your designated organization name at VividCortex.

We support three types of custom metrics:

  • counter (‘c’): a value that increases or stays the same overtime.

  • gauge (‘g’): which is just some current measured value, like a server variable or motherboard temperature.

  • derivative (’d’): which takes the currently seen value for each new sample, subtracts it by the previously seen value and then divides the result by the time between samples. For instance, if you have some value that was always increasing, like “total process time”, a derivative will compute a per-second throughput, rather than having a metric that is just getting bigger and bigger always.

You can search and visualize your custom metrics alongside all others in the Metrics dashboard. Here’s how a test custom metric we created named custom.mytable.auto_increment looks like in the dashboard, showing a constant increase in the AUTO_INCREMENT column as successive inserts were being done in this table at a rate of 2 per second:

Custom metric showing in the Metrics dashboard